You are viewing limited content. For full access, please sign in.

Question

Question

Is it possible to achieve Laserfiche high availability without using shared disk?

Laserfiche How To Ideas Self-Hosted
Updated August 20
asked on July 31, 2023

Hi,

 

We have a customer asking how to achieve Laserfiche high availability without using shared disk?

 

As we know Laserfiche is supporting failover cluster which requires a disk to be available for all the nodes, so the cluster will switch it along with the services to the passive node in case if current active node failure.

 

In the other hand the SQL server support Always on which not using shared disk.

 

Is there any method or techniques  which Laserfiche support to achieve this scenario? 

 

0 0

Answer

APPROVED ANSWER SELECTED ANSWER
replied on July 31, 2023

Hi Abdellah,

Yes, you can use highly available network storage for your data rather than a shared disk. This way the storage holding your Laserfiche data doesn't have to be part of the failover cluster configuration itself.

For example, a highly available Windows File Server cluster, Scale-Out File Server, SMB-compatible storage appliance like NetApp, or file storage service like Azure Files or Amazon FSx for Windows File Server. In this scenario, your repository data (volumes, audit logs, search catalogs, etc.) is pointed at a UNC path (\\fileserver\laserfiche\repository) or mapped network drive for that storage location.

Note that any HA data replication is handled transparently at the storage layer. The Laserfiche application layer does not handle any data replication, unlike SQL Server Always-On Availability Groups.

2 0
View 5 previous replies
replied on August 2, 2023

Hi Samuel,

 

Thank you for quick and valuable reply.

 

So, for confirmation there is no way to achieve this with their vSAN storage? Because they don't have the highly available network storage solutions mentioned  above. 

0 0
replied on August 2, 2023

To be more obvious they are using VMware infrastructure (VMware SDDC), which using vSAN storage. 

 

  1. So, using shared disk on VMware infrastructure causes many limitations such as Snapshots
  2. Live (online) storage vMotion
  3. Hot add memory/vCPU .

 

Therefor they are looking for alternative approaches to achieve the HA for ECM servers without using shared disk.

 

0 0
replied on August 2, 2023 • Show version history

So, for confirmation there is no way to achieve this with their vSAN storage? Because they don't have the highly available network storage solutions mentioned  above. 

I didn't say that - my list above was only some examples.

vSAN File Share with an SMB File Share looks like it could be an option as it's a "SMB-compatible file share/storage appliance". The customer should thoroughly research vSAN File Share to determine if it can meet their needs or not. 

This looks like a good walkthrough: vSAN File Share with vSAN File Service Configuration and Setup

1 0
replied on August 3, 2023 • Show version history

Abdellah, another thought. The customer should consider if they actually need Windows Server Failover Clustering (WSFC) for high availability. It's complicated to configure (as you're finding out) and virtualization-level solutions like VMware vSphere High Availability can often achieve the same practical result in a much easier way.

Only Laserfiche Server and Laserfiche Directory Server support Windows failover clustering; other components like Forms, Workflow, etc. do not.

VMware vSphere High Availability can protect everything.

1 0
replied on August 5, 2023

Great! 

So VMware vSphere HA recommended and easy to configure. I'm wondering how Laserfiche is supporting this solution or how to deploy it, is it just a regular installation of LF modules (like LF Server, LFDS) on virtual machines and then cloning these VMs as a secondary VMS on different hosts and VMware vSphere HA will handle every thing between them?

 

Is there any special configuration or consideration on Laserfiche side or just a normal install/config like on a single machines?

0 0
replied on August 14, 2023

The great thing about the VMware HA approach is that there is nothing different at all on the Laserfiche side. 

is it just a regular installation of LF modules (like LF Server, LFDS) on virtual machines and then cloning these VMs as a secondary VMS on different hosts and VMware vSphere HA will handle every thing between them?

Exactly.

Is there any special configuration or consideration on Laserfiche side or just a normal install/config like on a single machines?

No special configuration. Treat it like a normal install/config.

1 0
replied on August 20

Hi Samuel,


We are planning to deploy our entire Laserfiche 11 environment on Alibaba Cloud and need guidance on the recommended supported High Availability architecture.

Our current Alibaba Cloud environment does not support SCSI-3 Persistent Reservations. According to Alibaba Cloud Support, there is currently no shared storage product available in the current Alibaba Cloud region hosting our environment that meets the requirements for WSFC with Shared Storage. Therefore, we cannot use the traditional WSFC/Shared Storage approach.

For the database layer, we are considering ApsaraDB RDS for Microsoft SQL Server 2022 Standard Edition, using High-Availability Edition with Automatic Failover.

Our main questions are:

  1. Is ApsaraDB RDS for Microsoft SQL Server, with High Availability and Automatic Failover, a supported configuration for Laserfiche 11?
  2. If the SQL database is highly available through RDS, can Laserfiche 11 be deployed without WSFC and without Shared Storage?
  3. What is the recommended supported architecture for achieving High Availability for the Laserfiche 11 application and Repository data in this scenario?
  4. Specifically, how should the Laserfiche Repository volumes and data be handled to provide failover and avoid a single point of failure when shared storage is not available?
  5. Are there any specific requirements or limitations for deploying Laserfiche 11 application components on Alibaba Cloud VMs in this HA architecture?

Our objective is to implement a fully supported Laserfiche 11 High Availability solution on Alibaba Cloud without WSFC or Shared Storage.

We would appreciate Laserfiche's recommended architecture for this scenario.

 

0 0
replied on August 20

Hello Mahmoud,

First, if this is a new deployment, I strongly recommend you upgrade to Laserfiche 12 first and deploy that instead of Laserfiche 11. At this point, Laserfiche 12 has hundreds to thousands of bug fixes (including security fixes) over Laserfiche 11.

To your specific questions:

  1. Laserfiche supports Microsoft SQL Server's official HA/failover mechanisms, Failover Cluster Instances (FCI) and Always On availability groups. If ApsaraDB RDS for SQL Server is using one or both of those for its "High-availability Edition", those database HA configurations are supported. You should confirm with Alibaba Cloud support.
    1. Laserfiche does not specifically test with the ApsaraDB RDS for SQL Server managed service, so I cannot give you an official statement that that exact service is "supported". Laserfiche supports SQL Server. If Alibaba's managed service for SQL Server runs it in a configuration Microsoft supports, there should be no issues.
  2. Database HA and application server/storage HA are separate questions. 
  3. If you don't have shared storage available, you have one option for application/storage-level HA: a continuous or async replication disaster recovery solution between zones. Alibaba's offering for that is EBS Async Replication vs CDR for ECS Disaster Recovery - Cloud Backup - Alibaba Cloud - ECS disaster recovery - continuous replication | Cloud Backup - Alibaba Cloud Documentation Center - Cross-region and cross-zone disaster recovery-ECS - Cloud Backup - Alibaba Cloud. Note that CDR may not be available for current versions of Windows Server per their OS support matrix. They say async replication supports VM consistency groups, which is good. You would add all Laserfiche VMs, so they have the same recovery point.

    When there are HA requirements, I generally prefer continuous replication architectures over WSFC/shared storage ones, even when those options are available. WSFC adds significant complexity to the application-layer configuration and there are many ways things can go wrong with that added complexity. Continuous replication happens below the application layer and does not add that complexity. In my opinion, it is usually safer and lower risk.
    1. Laserfiche does not provide official support statements for specific backup and BC/DR solutions. They are infrastructure solutions that work at a level below Laserfiche. However, many Laserfiche customers use continuous replication solutions to protect their production Laserfiche systems.
  4. See response to #3 above.
  5. No specific requirements with a continuous/async replication HA approach - that's one of the main benefits. For failover, you need to ensure the databases are restored to the same point as the VM/storage checkpoint, so the databases, volumes, and application state stays synchronized.
0 0

Replies

You are not allowed to reply in this post.
You are not allowed to follow up in this post.

Sign in to reply to this post.

Details
Related Posts
Loading ...